Underestimates of intussusception rates among US infants based on inpatient discharge data: implications for monitoring the safety of rotavirus vaccines.

BACKGROUND: Because a previous rotavirus vaccine was associated with intussusception, new rotavirus vaccines are monitored postlicensure for any such association. Accurate background intussusception rates are needed to determine whether the number of cases observed after vaccination exceeds that expected by chance. Previously, intussusception rates were obtained from inpatient discharge databases. We sought to determine the rate of intussusception among infants managed only with short-stay or emergency department care.
METHODS: Intussusception cases occurring in infants were identified retrospectively at 3 children's hospitals from January 2001 through March 2006, a period without rotavirus vaccine use, by a search of discharge, billing, and radiology databases for International Classification of Diseases, Ninth Revision, Clinical Modification code 560.0 (intussusception) and procedure codes and by review of medical records.
RESULTS: Of 156 infants with intussusception fulfilling Brighton level 1 criteria, 81 (52%) were billed as inpatients, 68 (44%) as short-stay patients, and 7 (4%) as emergency department patients only. The use of only inpatients assigned code 560.0 underestimated the total number of level 1 cases at the hospitals by 44%. The mean annual intussusception rate for the hospitals' catchment counties was 49.3 cases per 100,000 live births (inpatient cases: 27.1 cases per 100,000 live births; short-stay or emergency department cases: 22.3 cases per 100,000 live births).
CONCLUSIONS: Intussusception rates based solely on inpatient discharge databases could underestimate the true incidence of level 1 intussusception by >40%. Background rates used for assessment of risk after vaccination should account for cases managed only with short-stay or emergency department care.